Re: GNUstep for Darwin (radar bug 2783924 / gcc 3.x affection)

[LHelldorf]

>LHelldorf@kabel.de wrote:
>> 
>> >Adam Fedor wrote:
>> >>
>> >> GNUstep is close to being ported. Right now there is an odd bug in
>> >> Darwin GCC that causes the compiler to apparently output Altivec
>> >> instructions for some GCC __builtin functions even on machines that
>> >> don't have Altivec
>> >
>> >Do I know about this one already?  I thought we'd eliminated all
>> >the accidental appearances of AltiVec instructions a while ago.
>> >
>> >Stan
>> Hi Stan,
>> I shouted about this bug a while ago (09/24/2001 till 10/02/2001, 
several
>> messages) on darwin-development@lists.apple.com. Aren't you on that 
list anymore?
>
>OK, I did see that one, just didn't connect it to GNUstep problems.
>(I get 600 msgs/day typically, easy to overlook one!)
>
>> Anyway, the radar bug id is 2783924 and there has not been much done on 
this bug since <GMT12-Oct-2001 20:42:50GMT> as a look at 
>http://www.opensource.apple.com/bugs/numericIndex.html tells me. It would be 
great to have this bug fixed soon since it is
>> hindering porting GNUstep to Darwin.
>
>Most everybody in the compiler group is now working on 3.x in
>preparation for switching over, so there's not a lot of cycles
>available for 2.95 bugs.  I'm sorry I don't have any better news
>on this one.

gcc 3.x would be fine for me if it comes with the standard install of 
Darwin/Mac OS X. Infact every compiler would be fine for me if it works 
and if the user has not to fiddle around building the compiler and stuff. 
I prefer the easiest posible way to install GNUstep (GNUstep installation 
is already complicated enough), and this way doesn't include building the 
compiler.

Before I forget this: GCC 3.X IS AFFECTED BY THAT BUG TOO! (according to 
http://www.opensource.apple.com/bugs/numericIndex.html, didn't find out 
that myself since I ain't got gcc 3.x)

>
>Stan

Greetings, Lars